apache-airflow-providers-apache-hive 9.0.6__py3-none-any.whl → 9.1.0__py3-none-any.whl

airflow/providers/apache/hive/__init__.py

@@ -29,11 +29,11 @@ from airflow import __version__ as airflow_version
 
 __all__ = ["__version__"]
 
-__version__ = "9.0.6"
+__version__ = "9.1.0"
 
 if packaging.version.parse(packaging.version.parse(airflow_version).base_version) < packaging.version.parse(
-    "2.9.0"
+    "2.10.0"
 ):
     raise RuntimeError(
-        f"The package `apache-airflow-providers-apache-hive:{__version__}` needs Apache Airflow 2.9.0+"
+        f"The package `apache-airflow-providers-apache-hive:{__version__}` needs Apache Airflow 2.10.0+"
     )

airflow/providers/apache/hive/hooks/hive.py

@@ -27,13 +27,17 @@ from collections.abc import Iterable, Mapping
 from tempfile import NamedTemporaryFile, TemporaryDirectory
 from typing import TYPE_CHECKING, Any
 
+from deprecated import deprecated
+from typing_extensions import Literal, overload
+
 if TYPE_CHECKING:
     import pandas as pd
+    import polars as pl
 
 import csv
 
 from airflow.configuration import conf
-from airflow.exceptions import AirflowException
+from airflow.exceptions import AirflowException, AirflowProviderDeprecationWarning
 from airflow.hooks.base import BaseHook
 from airflow.providers.common.compat.version_compat import AIRFLOW_V_3_0_PLUS
 from airflow.providers.common.sql.hooks.sql import DbApiHook
@@ -1031,37 +1035,106 @@ class HiveServer2Hook(DbApiHook):
         schema = kwargs["schema"] if "schema" in kwargs else "default"
         return self.get_results(sql, schema=schema, hive_conf=parameters)["data"]
 
-    def get_pandas_df(  # type: ignore
+    def _get_pandas_df(  # type: ignore
         self,
         sql: str,
         schema: str = "default",
         hive_conf: dict[Any, Any] | None = None,
         **kwargs,
     ) -> pd.DataFrame:
+        try:
+            import pandas as pd
+        except ImportError as e:
+            from airflow.exceptions import AirflowOptionalProviderFeatureException
+
+            raise AirflowOptionalProviderFeatureException(e)
+
+        res = self.get_results(sql, schema=schema, hive_conf=hive_conf)
+        df = pd.DataFrame(res["data"], columns=[c[0] for c in res["header"]], **kwargs)
+        return df
+
+    def _get_polars_df(  # type: ignore
+        self,
+        sql: str,
+        schema: str = "default",
+        hive_conf: dict[Any, Any] | None = None,
+        **kwargs,
+    ) -> pl.DataFrame:
+        try:
+            import polars as pl
+        except ImportError as e:
+            from airflow.exceptions import AirflowOptionalProviderFeatureException
+
+            raise AirflowOptionalProviderFeatureException(e)
+
+        res = self.get_results(sql, schema=schema, hive_conf=hive_conf)
+        df = pl.DataFrame(res["data"], schema=[c[0] for c in res["header"]], orient="row", **kwargs)
+        return df
+
+    @overload  # type: ignore[override]
+    def get_df(
+        self,
+        sql: str,
+        schema: str = "default",
+        hive_conf: dict[Any, Any] | None = None,
+        *,
+        df_type: Literal["pandas"] = "pandas",
+        **kwargs: Any,
+    ) -> pd.DataFrame: ...
+
+    @overload  # type: ignore[override]
+    def get_df(
+        self,
+        sql: str,
+        schema: str = "default",
+        hive_conf: dict[Any, Any] | None = None,
+        *,
+        df_type: Literal["polars"],
+        **kwargs: Any,
+    ) -> pl.DataFrame: ...
+
+    def get_df(  # type: ignore
+        self,
+        sql: str,
+        schema: str = "default",
+        hive_conf: dict[Any, Any] | None = None,
+        *,
+        df_type: Literal["pandas", "polars"] = "pandas",
+        **kwargs,
+    ) -> pd.DataFrame | pl.DataFrame:
         """
-        Get a pandas dataframe from a Hive query.
+        Get a pandas / polars dataframe from a Hive query.
 
         :param sql: hql to be executed.
         :param schema: target schema, default to 'default'.
         :param hive_conf: hive_conf to execute alone with the hql.
+        :param df_type: type of dataframe to return, either 'pandas' or 'polars'
         :param kwargs: (optional) passed into pandas.DataFrame constructor
         :return: result of hive execution
 
         >>> hh = HiveServer2Hook()
         >>> sql = "SELECT * FROM airflow.static_babynames LIMIT 100"
-        >>> df = hh.get_pandas_df(sql)
+        >>> df = hh.get_df(sql, df_type="pandas")
         >>> len(df.index)
         100
 
-        :return: pandas.DateFrame
+        :return: pandas.DateFrame | polars.DataFrame
         """
-        try:
-            import pandas as pd
-        except ImportError as e:
-            from airflow.exceptions import AirflowOptionalProviderFeatureException
-
-            raise AirflowOptionalProviderFeatureException(e)
-
-        res = self.get_results(sql, schema=schema, hive_conf=hive_conf)
-        df = pd.DataFrame(res["data"], columns=[c[0] for c in res["header"]], **kwargs)
-        return df
+        if df_type == "pandas":
+            return self._get_pandas_df(sql, schema=schema, hive_conf=hive_conf, **kwargs)
+        if df_type == "polars":
+            return self._get_polars_df(sql, schema=schema, hive_conf=hive_conf, **kwargs)
+
+    @deprecated(
+        reason="Replaced by function `get_df`.",
+        category=AirflowProviderDeprecationWarning,
+        action="ignore",
+    )
+    def get_pandas_df(  # type: ignore
+        self,
+        sql: str,
+        schema: str = "default",
+        hive_conf: dict[Any, Any] | None = None,
+        **kwargs,
+    ) -> pd.DataFrame:
+        return self._get_pandas_df(sql, schema=schema, hive_conf=hive_conf, **kwargs)

{apache_airflow_providers_apache_hive-9.0.6.dist-info → apache_airflow_providers_apache_hive-9.1.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: apache-airflow-providers-apache-hive
-Version: 9.0.6
+Version: 9.1.0
 Summary: Provider package apache-airflow-providers-apache-hive for Apache Airflow
 Keywords: airflow-provider,apache.hive,airflow,integration
 Author-email: Apache Software Foundation <dev@airflow.apache.org>
@@ -20,10 +20,9 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: System :: Monitoring
-Requires-Dist: apache-airflow>=2.9.0
-Requires-Dist: apache-airflow-providers-common-sql>=1.20.0
+Requires-Dist: apache-airflow>=2.10.0
+Requires-Dist: apache-airflow-providers-common-sql>=1.26.0
 Requires-Dist: hmsclient>=0.1.0
-Requires-Dist: pandas>=2.1.2,<2.2
 Requires-Dist: pyhive[hive-pure-sasl]>=0.7.0
 Requires-Dist: thrift>=0.11.0
 Requires-Dist: jmespath>=0.7.0
@@ -35,8 +34,8 @@ Requires-Dist: apache-airflow-providers-presto ; extra == "presto"
 Requires-Dist: apache-airflow-providers-samba ; extra == "samba"
 Requires-Dist: apache-airflow-providers-vertica ; extra == "vertica"
 Project-URL: Bug Tracker, https://github.com/apache/airflow/issues
-Project-URL: Changelog, https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.0.6/changelog.html
-Project-URL: Documentation, https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.0.6
+Project-URL: Changelog, https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.1.0/changelog.html
+Project-URL: Documentation, https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.1.0
 Project-URL: Mastodon, https://fosstodon.org/@airflow
 Project-URL: Slack Chat, https://s.apache.org/airflow-slack
 Project-URL: Source Code, https://github.com/apache/airflow
@@ -74,7 +73,7 @@ Provides-Extra: vertica
 
 Package ``apache-airflow-providers-apache-hive``
 
-Release: ``9.0.6``
+Release: ``9.1.0``
 
 
 `Apache Hive <https://hive.apache.org/>`__
@@ -87,7 +86,7 @@ This is a provider package for ``apache.hive`` provider. All classes for this pr
 are in ``airflow.providers.apache.hive`` python package.
 
 You can find package information and changelog for the provider
-in the `documentation <https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.0.6/>`_.
+in the `documentation <https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.1.0/>`_.
 
 Installation
 ------------
@@ -104,10 +103,9 @@ Requirements
 =======================================  ==================
 PIP package                              Version required
 =======================================  ==================
-``apache-airflow``                       ``>=2.9.0``
-``apache-airflow-providers-common-sql``  ``>=1.20.0``
+``apache-airflow``                       ``>=2.10.0``
+``apache-airflow-providers-common-sql``  ``>=1.26.0``
 ``hmsclient``                            ``>=0.1.0``
-``pandas``                               ``>=2.1.2,<2.2``
 ``pyhive[hive_pure_sasl]``               ``>=0.7.0``
 ``thrift``                               ``>=0.11.0``
 ``jmespath``                             ``>=0.7.0``
@@ -140,5 +138,5 @@ Dependent package
 ======================================================================================================================  ===================
 
 The changelog for the provider package can be found in the
-`changelog <https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.0.6/changelog.html>`_.
+`changelog <https://airflow.apache.org/docs/apache-airflow-providers-apache-hive/9.1.0/changelog.html>`_.
 

{apache_airflow_providers_apache_hive-9.0.6.dist-info → apache_airflow_providers_apache_hive-9.1.0.dist-info}/RECORD

@@ -1,8 +1,8 @@
 airflow/providers/apache/hive/LICENSE,sha256=gXPVwptPlW1TJ4HSuG5OMPg-a3h43OGMkZRR1rpwfJA,10850
-airflow/providers/apache/hive/__init__.py,sha256=vNoiAoYLF_9C7nqGSP4r3R85TLGsvMfUPet08HE40YM,1498
+airflow/providers/apache/hive/__init__.py,sha256=4A8vByNcISGrsZVhWBPVTcf7lp2FsDeWev2kifiVVxM,1500
 airflow/providers/apache/hive/get_provider_info.py,sha256=PdxdiZ7fdK1D-BTyZqj1oMS_1GS8_kbz2hZVgWMTkFI,5569
 airflow/providers/apache/hive/hooks/__init__.py,sha256=mlJxuZLkd5x-iq2SBwD3mvRQpt3YR7wjz_nceyF1IaI,787
-airflow/providers/apache/hive/hooks/hive.py,sha256=GZYlNhwI0HF9RqTzU3X7BBs-TwVqndoXTEvcT5Yphdw,42779
+airflow/providers/apache/hive/hooks/hive.py,sha256=ay_4PN2DI4NDfk4z88OgbQTeQ5_FlajBZVoYaaEvUhM,45145
 airflow/providers/apache/hive/macros/__init__.py,sha256=mlJxuZLkd5x-iq2SBwD3mvRQpt3YR7wjz_nceyF1IaI,787
 airflow/providers/apache/hive/macros/hive.py,sha256=FEMCR9rEWKFOKtKAvhVWhD5jRJnSYeHepVGgJLzUa2k,4548
 airflow/providers/apache/hive/operators/__init__.py,sha256=mlJxuZLkd5x-iq2SBwD3mvRQpt3YR7wjz_nceyF1IaI,787
@@ -21,7 +21,7 @@ airflow/providers/apache/hive/transfers/mssql_to_hive.py,sha256=9NVgJESAwPvxl3D1
 airflow/providers/apache/hive/transfers/mysql_to_hive.py,sha256=6BkAoYrn9wDfSC0BKHp86EgoUVtqwHGz9nZsielCydQ,7022
 airflow/providers/apache/hive/transfers/s3_to_hive.py,sha256=UDmTXp1mMNWl31qBe_Z3Sqp7VLJym6kQk3gsxkAxP3g,12417
 airflow/providers/apache/hive/transfers/vertica_to_hive.py,sha256=IBiMGF7322rOrwDbWMVB89oKO8hFhhBmc6rQwPkFA34,5555
-apache_airflow_providers_apache_hive-9.0.6.dist-info/entry_points.txt,sha256=Hzixt33mYYldwmwswarArUB7ZU0xbmUtd3tFViZ414s,185
-apache_airflow_providers_apache_hive-9.0.6.dist-info/WHEEL,sha256=G2gURzTEtmeR8nrdXUJfNiB3VYVxigPQ-bEQujpNiNs,82
-apache_airflow_providers_apache_hive-9.0.6.dist-info/METADATA,sha256=rcSnAiHtHCbTln0U1bEn-AW6oJHI9XxtBqbD-ijE0Ns,7194
-apache_airflow_providers_apache_hive-9.0.6.dist-info/RECORD,,
+apache_airflow_providers_apache_hive-9.1.0.dist-info/entry_points.txt,sha256=Hzixt33mYYldwmwswarArUB7ZU0xbmUtd3tFViZ414s,185
+apache_airflow_providers_apache_hive-9.1.0.dist-info/WHEEL,sha256=G2gURzTEtmeR8nrdXUJfNiB3VYVxigPQ-bEQujpNiNs,82
+apache_airflow_providers_apache_hive-9.1.0.dist-info/METADATA,sha256=RhQwm3ChWLCvmAuQOh5OetppPh8LVWOgxJR0VjZ3mNs,7104
+apache_airflow_providers_apache_hive-9.1.0.dist-info/RECORD,,